agent-native-architecture

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs agents to fetch and process open web content and user-generated messages (e.g., references/architecture-patterns.md shows ResearchAgent with WebTools.webSearch/webFetch, references/mcp-tool-design.md defines a call_api that fetches arbitrary URLs, and the Event-Driven Agent example processes Discord messages), so the agent will ingest untrusted third‑party content that can influence subsequent tool use and decisions.